基于神经网络的离网自适应光储一体化发电系统
Off Grid Adaptive Optical Storage Integrated Power Generation System Based on Neural Network
西藏电网网架结构薄弱,供电半径大,加上高渗透率可再生能源的并网造成的调控不确定性,使得对偏远地区的供电极不稳定。为实现对偏远用户的平稳供电,提出基于神经网络的离网自适应光储一体化发电系统。首先,构建了光储一体化发电系统模型,并提出基于神经网络(NN)训练的最大功率点跟踪(MPPT)算法,提高系统在天气突变时仍以最大功率输出电能的能力。然后,提出了混合储能模式,建立基于DC-DC变换器的超级电容控制策略和储能电池的充放电控制策略,减小光伏发电的不确定性。最后,设置了两种情景对比验证所提系统的有效性,仿真结果表明,基于神经网络的离网自适应光储一体化发电系统相较于采用传统电导增量法MPPT的发电系统,其电压波动性更小,电能质量更高。设计过程和方法为后续系统设计提供了一些经验借鉴。
The power grid in Tibet suffers from a weak infrastructure and extensive power supply radius. Coupled with the regulatory uncertainty resulting from the integration of highly penetrative renewable energy sources, the electricity supply to remote areas remains highly unstable. In order to realize stable power supply for remote users, an off-grid adaptive optical storage integrated power generation system based on neural network was proposed. Firstly, the model of integrated optical storage power generation system was established, and the maximum power point tracking (MPPT) algorithm based on neural network (NN) training was introduced. This algorithm enhanced the system's capacity to deliver maximum power output even during sudden weather changes. Next, the hybrid energy storage mode was proposed, wherein control strategies were developed for both the supercapacitor, employing a DC-DC converter and the energy storage battery, aiming to mitigate uncertainties associated with photovoltaic power generation. Subsequently, the effectiveness of the proposed system through a comparative analysis of two scenarios was validated. The simulation results demonstrated that the off-grid adaptive optical storage integrated power generation system based on neural network exhibited the reduced voltage fluctuation and higher power quality compared to traditional MPPT-based power generation systems using the conductance increment method. The design process and methodologies presented herein offered valuable insights for subsequent system design efforts.
integration of photovoltaic and energy storage / neural network / mixed energy storage / off grid
| [1] |
王岩,李鹏,唐劲飞 |
| [2] |
|
| [3] |
姚兴佳,刘喜,朱家玲 |
| [4] |
闵江威.光伏发电系统的最大功率点跟踪控制技术研究[D].武汉: 华中科技大学,2006. |
| [5] |
刘春霞,刘立群.免疫小脑模型神经网络优化光伏输出[J].电力系统保护与控制 2011,39(11):145-148 |
| [6] |
|
| [7] |
朱达祥, 群体智能优化算法—粒子群算法的研究和改进[D]. 无锡: 江南大学, 2017. |
| [8] |
陈明. MATLAB神经网络原理与实例精解[M].北京:清华大学出版社,2013. |
| [9] |
张瑞林.人工神经网络在纺织中的应用研究[D].杭州: 浙江大学,2001. |
| [10] |
张剑湖,叶峰.人工神经网络的模型,特征及其发展方向[J].现代电子技术,2004(12):57-60. |
| [11] |
张巧超,曾昭冰.人工神经网络概述[J].辽宁经济管理干部学院(辽宁经济职业技术学院学报),2010(04):68-69. |
西藏农牧学院大学生创新创业训练项目(533322012)
/
| 〈 |
|
〉 |